subChains.twDEMC: subChains twDEMC

In twDEMC: parallel DEMC

Description

Condenses an twDEMC List to the chains iChains e.g 1:4.

Usage

## S3 method for class 'twDEMC'
subChains(x, iChains = NULL, iPops = NULL, nPop = getNPops(x), 
    ..., doKeepBatchCall = FALSE)

Arguments

x	the twDEMC object to select chains from
iChains	integer vector: the chains to select
iPops	integer vector: the populations, whose chains should be selected
nPop	number of populations in x
...	not used
doKeepBatchCall	wheter to retain the batch call attribute of x

Details

Alternatively to specification of iChains, one can specify a vector of populations and the total number of populations.

The twDEMC class

The twDEMC class represents MCMC results with all chains having the same length. This is usually a result by a call to concatPops.twDEMCPops. It holds the same results as the twDEMCPops class, except that the entries in pops list are concatenated by chain.

• parms: array (nStep, nParm, nChain) of parameter samples

• temp: matrix (nStep, nResComp ) of temperature, one collumn for each logDensity components

• parms: array (nStep, nBlock, nChain) acceptance rate

• resLogDen: array (nStep, nResComp, nChain) results of logDensities

• logDen: array (nStep, nBlock, nChain) sum across logDensity components for each block

• Y: array (lastUnthinnedSteps, nParm+nBlock+nResComp, nChain) of parameters, acceptance per block and result components of latest unthinned steps

There are several methods access properties, i.e. an object of class twDEMC

• number of generations: getNGen.twDEMC

• number of samples (only one sample each thinning inteval): getNSamples.twDEMC

• number of chains: getNChains.twDEMC

• number of populations: getNPops.twDEMC

• number of chains per population: getNChainsPop.twDEMC

• number of parameters: getNParms.twDEMC

• thinning interval: res$thin

There are several methods to transform or subset:

• select chains or sub-populations: this method

• thin all the chains: thin.twDEMC

• select subset of cases: subset.twDEMC

• combine several twDEMC results to a bigger set of populations combinePops.twDEMC

• stack all the results of all chains to one big matrix stackChains.twDEMC

There are several methods utilize the functions of the coda package.

• convert an twDEMC to a coda mcmc.list as.mcmc.list.twDEMC

• applying a function to all of the chains: mcmcListApply

• stack all the results of all chains to one big matrix: stackChains.mcmc.list

• transforming parameters transOrigPopt.mcmc.list

Value

A list of class twDEMC (see twDEMCBlockInt).

Author(s)

Thomas Wutzler

See Also

twDEMC

twDEMC documentation built on May 2, 2019, 5:38 p.m.